New method for true-triaxial rock testing

Abstract

Sub-surface structures provide attractive alternatives for storage of explosives and other military hardware. These facilities are commonly constructed at shallow depths where rocks have undergone extensive weathering and the geologic system contains joints and discontinuities. For underground munitions storage structures, risk and performance assessment studies have to be conducted to qualify the site and establish the “clear zone” in case of accidental detonation. At high loading densities, accidental detonation of a munitions storage facility can lead to rupturing of the overburden cover and creation of a hazardous fragment environment. The degree and extent of the “clear zone” is controlled by the overall characteristics of the geologic and engineered systems. Rock joint spacing is considered to be important in prediction of the hazardous range of blast-induced fragments and associated impact energy. A full scale tunnel explosion test was conducted in 1988 in the desert of California. The tunnel was constructed in a weathered, jointed, igneous rock mass with several, well defined discontinuities. The information from the overall characteristics of this tunnel was used to construct five scaled model tunnels, with simulated joints and discontinuities, under physical modeling at 1-g. The loading densities used in two of the model tests were equivalent to that of the full scale event. For the other three tests, loading densities were changed in order to determine the effects of the explosive weight on the jointed rockmass response. The tests reported in this paper are unique in terms of size and simulation method. Five large test beds were constructed in trenches filled with model material simulating the full scale jointed rockmass with through-going discontinuities at 20:1 scale. Geometric and strength related properties were scaled, whereas the density scale factors were maintained at unity. The model materials were formulated in such a way that similitude conditions between the properties of the full scale rock mass were maintained. Test beds were constructed by step-by-step casting of the model materials into the excavated trenches. Impedance characteristics of the model materials were matched with those of the host-ground for realistic simulation of the ground shock propagation. This paper provides a brief discussion on the tests performed and elaborates on the applicability and the economics of the physical modeling for studies related to explosives-underground-structures interaction.